From: Rik van Riel <riel@redhat.com>
To: "Huang, Ying" <ying.huang@intel.com>,
Andrew Morton <akpm@linux-foundation.org>
Cc: tim.c.chen@intel.com, dave.hansen@intel.com,
andi.kleen@intel.com, aaron.lu@intel.com, linux-mm@kvack.org,
linux-kernel@vger.kernel.org, Hugh Dickins <hughd@google.com>,
Shaohua Li <shli@kernel.org>, Minchan Kim <minchan@kernel.org>,
Mel Gorman <mgorman@techsingularity.net>,
Tejun Heo <tj@kernel.org>, Wu Fengguang <fengguang.wu@intel.com>
Subject: Re: [PATCH] mm: Don't use radix tree writeback tags for pages in swap cache
Date: Thu, 25 Aug 2016 15:44:03 -0400 [thread overview]
Message-ID: <1472154243.2751.44.camel@redhat.com> (raw)
In-Reply-To: <1472153230-14766-1-git-send-email-ying.huang@intel.com>
On Thu, 2016-08-25 at 12:27 -0700, Huang, Ying wrote:
> File pages use a set of radix tags (DIRTY, TOWRITE, WRITEBACK, etc.)
> to
> accelerate finding the pages with a specific tag in the radix tree
> during inode writeback.A A But for anonymous pages in the swap cache,
> there is no inode writeback.A A So there is no need to find the
> pages with some writeback tags in the radix tree.A A It is not
> necessary
> to touch radix tree writeback tags for pages in the swap cache.
>
> With this patch, the swap out bandwidth improved 22.3% (from ~1.2GB/s
> to
> ~ 1.48GBps) in the vm-scalability swap-w-seq test case with 8
> processes.
> The test is done on a Xeon E5 v3 system.A A The swap device used is a
> RAM
> simulated PMEM (persistent memory) device.A A The improvement comes
> from
> the reduced contention on the swap cache radix tree lock.A A To test
> sequential swapping out, the test case uses 8 processes, which
> sequentially allocate and write to the anonymous pages until RAM and
> part of the swap device is used up.
>
> Details of comparison is as follow,
>
> baseA A A A A A A A A A A A A base+patch
> ---------------- --------------------------
> A A A A A A A A A %stddevA A A A A %changeA A A A A A A A A %stddev
> A A A A A A A A A A A A A \A A A A A A A A A A |A A A A A A A A A A A A A A A A \
> A A A 1207402 A+-A A 7%A A A A A +22.3%A A A A 1476578 A+-A A 6%A A vmstat.swap.so
> A A A 2506952 A+-A A 2%A A A A A +28.1%A A A A 3212076 A+-A A 7%A A vm-
> scalability.throughput
> A A A A A 10.86 A+- 12%A A A A A -23.4%A A A A A A A 8.31 A+- 16%A A perf-profile.cycles-
> pp._raw_spin_lock_irq.__add_to_swap_cache.add_to_swap_cache.add_to_sw
> ap.shrink_page_list
> A A A A A 10.82 A+- 13%A A A A A -33.1%A A A A A A A 7.24 A+- 14%A A perf-profile.cycles-
> pp._raw_spin_lock_irqsave.__remove_mapping.shrink_page_list.shrink_in
> active_list.shrink_zone_memcg
> A A A A A 10.36 A+- 11%A A A A -100.0%A A A A A A A 0.00 A+- -1%A A perf-profile.cycles-
> pp._raw_spin_lock_irqsave.__test_set_page_writeback.bdev_write_page._
> _swap_writepage.swap_writepage
> A A A A A 10.52 A+- 12%A A A A -100.0%A A A A A A A 0.00 A+- -1%A A perf-profile.cycles-
> pp._raw_spin_lock_irqsave.test_clear_page_writeback.end_page_writebac
> k.page_endio.pmem_rw_page
>
> Cc: Hugh Dickins <hughd@google.com>
> Cc: Shaohua Li <shli@kernel.org>
> Cc: Minchan Kim <minchan@kernel.org>
> Cc: Rik van Riel <riel@redhat.com>
> Cc: Mel Gorman <mgorman@techsingularity.net>
> Cc: Tejun Heo <tj@kernel.org>
> Cc: Wu Fengguang <fengguang.wu@intel.com>
> Cc: Dave Hansen <dave.hansen@intel.com>
> Signed-off-by: "Huang, Ying" <ying.huang@intel.com>
> ---
> A mm/page-writeback.c | 6 ++++--
> A 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/mm/page-writeback.c b/mm/page-writeback.c
> index 82e7252..599d2f9 100644
> --- a/mm/page-writeback.c
> +++ b/mm/page-writeback.c
> @@ -2728,7 +2728,8 @@ int test_clear_page_writeback(struct page
> *page)
> A int ret;
> A
> A lock_page_memcg(page);
> - if (mapping) {
> + /* Pages in swap cache don't use writeback tags */
> + if (mapping && !PageSwapCache(page)) {
I wonder if that should be a mapping_uses_tags(mapping)
macro or similar, and a per-mapping flag?
I suspect there will be another case coming up soon
where we have a page cache radix tree, but no need
for dirty/writeback/... tags.
That use case would be DAX filesystems, where we do
use a struct page, but that struct page points at
persistent storage, and the tags are not necessary.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2016-08-25 19:44 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-08-25 19:27 [PATCH] mm: Don't use radix tree writeback tags for pages in swap cache Huang, Ying
2016-08-25 19:44 ` Rik van Riel [this message]
2016-08-29 19:47 ` Huang, Ying
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1472154243.2751.44.camel@redhat.com \
--to=riel@redhat.com \
--cc=aaron.lu@intel.com \
--cc=akpm@linux-foundation.org \
--cc=andi.kleen@intel.com \
--cc=dave.hansen@intel.com \
--cc=fengguang.wu@intel.com \
--cc=hughd@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mgorman@techsingularity.net \
--cc=minchan@kernel.org \
--cc=shli@kernel.org \
--cc=tim.c.chen@intel.com \
--cc=tj@kernel.org \
--cc=ying.huang@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).